Islamabad, 20 July: Trends Monitor, an investigative series being conducted by Media Matters for Democracy discovered a significant contribution of custom made bot platforms to top trending political hashtags in Pakistan. The study also suggests the presence of ‘networks’ of human-bots actively contributing to top political Twitter hashtags.
Trends Monitor, Media Matters for Democracy’s new investigative series has been launched to assess bot presence in top political hashtags on Twitter in Pakistan. Trends Monitor comprises of five weekly assessment reports to be launched over a period of two months. Weekly assessment reports comprise of an investigative assessment of top political hashtags trending in that week. The first [beta] report, i investigating top 3 political hashtags from 23rd to 30th June, launched on the 17th July includes assessment of #Election2018, #BBKaWaadaNibhanaHai and #اب_صرف_عمران_خان.
The research found the presence of custom-made platforms with bot-like attributes contributing with a large number of tweets.
“These [bot] platforms are specifically created to ‘automate’ the contribution to a hashtag for higher activity”, says Asad Baig, Founder, Executive Director of Media Matters for Democracy, “Bot platforms, as well call them, are different from simple ‘bot’ accounts. They are specific platforms, such as TweetDeck, or Twitter for iPhone and they are usually created for one purpose only; to automatically send a high volume of tweets to increase hashtag activity. This is classic bot-like behavior and hence the name bot-platforms”.
The assessment of 3 top political hashtags found bot platforms contributing to all three of the analysed hashtags. The highest activity of these bot-platforms was recorded in #اب_صرف_عمران_خان, where only 1 of them, named Staunch Insafian, contributed with 1% of the sample hashtag volume.
“1% contribution may not sound like much, but we need to understand that the hashtag volumes [number of tweets therein] can go as high as 133,100 tweets in just 7 days. In #اب_صرف_عمران_خان the recorded hashtag volume was 69,000 tweets. The sample that we used for analysis was 14,867 tweets and so 1% is 148 tweets. It’ll be much higher if we average that same figure [1%] for the total volume of #اب_صرف_عمران_خان, which is 69,000 tweets in the same period”, says Asad Baig.
The investigative study also ‘suggested’ the presence of, in most cases, organized human-bot networks contributing to these political hashtags. Human-bots are fake or dummy twitter accounts with a significantly high bot-like behavior but often manned by humans. Some of them were also found directly engaging in hateful and inciteful speech targeting political rivals, journalists and especially women.
“There is a clear trend that we see. We have found human-bots, in most cases, contributing to political hashtags to artificially populate them. In other cases, we also found them engaged in inciteful speech perhaps aimed at igniting political and religious violence against political rivals and journalists. It seems that there are different networks of them operating. Most human-bots in a network follow and retweet each other”, says Asad Baig.
Further analysing the top contributors of these hashtags, the investigative study finds that some of them have sent an ‘unnaturally high number of tweets’ to similar political hashtags supporting the narrative of one political party or another. Human-bots were also found attacking other human-bots posting tweets critical to the political leadership they seem to be supporting.
“It’s like a political cyber-war is ongoing on Twitter. It’s very obvious that accounts which have been excessively retweeting Maryam Nawaz Sharif’s and other PML-N outlets are attacking Imran Khan, and those tweeting in his favour. Similarly, accounts tweeting in favour of Imran Khan and PTI are attacking accounts tweeting in favour of PMLN and Maryam Nawaz herself. Inciteful speech is rampant in these exchanges”, explains Asad Baig.
Trends Monitor also found that the hateful and inciteful exchange wasn’t limited to political actors but was also often directed at journalists, especially women journalists.
“The hatespeech against women journalists and women politicians is often heavily sexualized. This is perhaps the most disturbing finding. We came across many extrmeley hateful, explicit and some inciteful items targeting journalists in a very organised manner. We are documenting what we can and report them to law enforcement”, says Sadaf Khan, Director Media Matters for Democracy.
Trends Monitor is based on an assessment Index developed by MMFD’s research team with an aim to detect unusual traffic in any given political hashtag. The Index also features indicators and identifiers to allow a qualitative analysis of top contributors — Twitter accounts which have sent the largest number of tweets and retweets.
The first full report is expected on 23rd July and will have an investigative assessment of political hashtags that trended between 2nd July to 8th July.
Other updates in this ongoing series will be reported.